Characterization of skin lesion texture in diffuse reflectance spectroscopic images

Abstract

This paper examines various texture features extracted from skin lesion images obtained by using diffuse reflectance spectroscopic imaging. Different image texture features have been applied to such images to separate precancerous from benign cases. These features are extracted based on the co-occurrence matrix, wavelet decomposition , fractal signature, and granulometric approaches. The results so far indicate that fractal and wavelet-based features are effective in distinguishing precancerous from benign cases.